Statistics

Critique Media Statistics

1

Is This a Good Claim?

Circle whether the evidence sounds strong or weak.

'9 out of 10 dentists recommend our toothpaste' — only 10 dentists asked

Strong
Weak

'Most Australians prefer our brand' — survey of 5,000 people across all states

Strong
Weak

'Our school is the best' — one student won a prize

Strong
Weak

'Crime has increased' — based on 10 years of police data

Strong
Weak
2

More Claim Evaluation

Circle whether the evidence sounds strong or weak.

'This diet works!' — tested on 3 people for 1 week

Strong
Weak

'Students are reading more' — based on 5 years of library data

Strong
Weak

'Everyone loves our product' — 100% approval in a company survey

Strong
Weak

'Exercise reduces stress' — study of 2,000 people over 2 years

Strong
Weak
3

Match the Problem

Match each statistical problem to its name.

Only 5 people surveyed
Graph starts at 95 not 0
Only one side shown
People pushed to answer a certain way
Biased question
Small sample size
Misleading graph
One-sided argument
4

Sort: Reliable or Unreliable Source?

Sort each source.

Government health statistics
One person's social media post
University research study
An advertisement
Census data
A friend's opinion
More Reliable
Less Reliable
5

What Makes a Survey Reliable?

Answer each question.

Why is a larger sample size better? ___

Why should a survey include diverse people? ___

What is a 'biased' question? Give an example: ___

6

Question the Claim

Write one question to check if each claim is reliable.

Claim: '80% of kids love our cereal!' Question: ___

Claim: 'Test scores have dropped this year.' Question: ___

Claim: 'This is the safest car on the road.' Question: ___

7

Spot the Problem

Circle the main problem.

'4 out of 5 kids prefer our drink' — company paid for survey

Small sample
Bias
Missing data

Graph shows sales 'doubled' but y-axis goes from 98 to 102

Small sample
Misleading scale
Missing data

'Average house price is $1.2M' — suburb has a few mansions

Outliers affect mean
Small sample
Bias
8

More Problem Spotting

Circle the main issue.

'90% agree!' — online poll anyone could take multiple times

Self-selection bias
Small sample
Misleading graph

Bar graph where one bar is wider than the other

Misleading visual
Small sample
Biased question

'Crime is skyrocketing' — based on one month

Cherry-picked data
Small sample
Misleading graph
9

Misleading Graphs

Explain how each technique can be misleading.

Starting the y-axis at a number other than 0: ___

Using pictures of different sizes instead of bars: ___

Leaving out data points that don't support the argument: ___

10

Rewrite the Claim

Rewrite each misleading claim to make it more honest.

Original: 'Everyone loves our product!' Better: ___

Original: 'Test scores have skyrocketed!' Better: ___

Original: '9 out of 10 recommend us!' Better: ___

11

Sample Size Matters

Circle the more reliable sample size for each survey.

Survey about favourite food:

Ask 5 friends
Ask 500 random people

Survey about exercise habits:

Survey 10 gym members
Survey 1,000 adults randomly

Testing a new medicine:

Test on 20 people
Test on 5,000 people

Student satisfaction survey:

Ask the top 3 students
Ask 100 random students
12

Sort: Biased or Unbiased Question?

Sort each survey question.

Don't you agree our school is the best?
How would you rate our school? (1-5)
Why do you love our cereal so much?
Which cereal do you prefer? (list options)
Do you support this obviously good policy?
What is your opinion on this policy?
Biased
Unbiased
13

Write Your Critique

Write a critique explaining why someone should be cautious.

Claim: 'Our app improved test scores by 50%!' Study tested 8 students over 2 weeks. Critique:

Claim: 'Eating chocolate makes you smarter!' Countries with more chocolate sales have more Nobel Prize winners. Critique:

14

Correlation vs Causation

Explain why correlation does not mean causation.

Ice cream sales and drowning rates both rise in summer. Does ice cream cause drowning? ___

Students who eat breakfast get better grades. Does breakfast cause better grades? ___

15

Design a Fair Survey

Design a fair survey about screen time.

Write 3 unbiased questions: ___

How many people would you survey and why? ___

How would you make your sample representative? ___

16

Home Activity: Media Detective

Find and critique statistics in the real world!

  • 1Find a statistic in a newspaper or online ad. Write down the claim.
  • 2Ask: How many people surveyed? Who paid? Is the graph fair?
  • 3Rewrite a misleading claim to make it more honest.
  • 4Create your own fair survey question about a topic you care about.
  • 5Find an ad using statistics. List 3 reasons it might be misleading.
17

Match the Graph Type to Its Bias

Match each graph feature to the type of bias it creates.

Truncated y-axis
3D bar chart
Missing labels
Cherry-picked time period
Perspective distorts bar height
Can't compare values
Makes small diff look large
Hides the long-term trend
18

Spot the Misleading Claim

Circle the option that correctly identifies the problem.

Graph shows y-axis from 98 to 102. Problem?

Makes 1% look like 100%
No problem
Axis too long

Survey of 5 people used to claim '80% of Australians'?

Sample too small
Math wrong
Good claim

Headline: 'Crime doubled!' (8 to 16 incidents). Problem?

Very small numbers
Should be tripled
No problem

Survey done by the company selling the product?

Biased source
More reliable
Fair survey
19

Sort: Fair vs Misleading

Sort each statement into the correct column.

Graph starts at zero
Survey of 1,000 random people
Y-axis starts at 95
Only positive results shown
Representative sample used
Poll paid for by the company
Fair
Misleading
20

Evaluate a Media Statistic

Read this claim and answer the questions below.

Claim: '9 out of 10 dentists recommend our toothpaste' — how many were surveyed? ___

What information is missing? ___

How would you verify this claim? ___

21

Track Misleading Techniques

Fill in the tally chart as you analyse these headlines.

ItemTallyTotal
Small sample
Biased source
Misleading graph
Correlation = causation
22

Rewrite the Headline

Rewrite each headline to make it more honest.

Original: 'Our supplement helps 4 in 5 users lose weight!' (study: 5 friends, 1 week). Rewrite: ___

Original: 'Coffee linked to success — successful people drink more coffee!' Rewrite: ___

Original: 'City crime up 200%!' (2 crimes last year, 6 this year). Rewrite: ___

23

Comparing Survey Sizes

This graph shows how many people were surveyed in each study. Each star = 100 people. Answer the questions.

Study A
Study B
Study C
Study D
Study E
1

Which study is most reliable? Why?

2

Which study should you be most sceptical of?

3

How many people total across all studies?

24

Sample Size and Reliability

Answer the questions about sample size.

Why does a larger sample give more reliable results? ___

Survey 1: 12 students. Survey 2: 500 students. Same result. Which do you trust more and why? ___

What is the minimum sample size you would consider reliable? ___

25

Order of Credibility

Number these sources 1 (most reliable) to 5 (least reliable).

?
Company's own website
?
Peer-reviewed journal
?
Anonymous social media post
?
Government health department
?
Friend's opinion
26

Graph Scale Investigation

Two graphs show the same data but look very different.

Graph A y-axis: 0–100. Graph B y-axis: 40–60. Same data. Which looks more dramatic? ___

Which graph is misleading and why? ___

How would you fix Graph B? ___

27

Choose the Better Question

Circle the more unbiased survey question.

Asking about diet:

Do you agree our product is the best?
Which product do you prefer?

Asking about school:

How many hours do you study?
Surely you study every night?

Asking about sport:

Most people love football, don't they?
Which sport do you enjoy most?

Asking about screen time:

How many hours per day do you use screens?
Are you addicted to screens?
28

Mean vs Median in the Media

A company claims 'Our workers earn an average of $90,000.'

Salaries: $30k, $35k, $40k, $45k, $390k. Mean = ___

Median = ___

Which is more representative? Why? ___

Why might the company choose to use the mean? ___

29

Classify the Evidence

Sort each type of evidence into the correct column.

Randomised controlled trial, n=10,000
One person's story
Government census data
Brand's own advertisement
Long-term longitudinal study
Poll of 8 friends
Strong evidence
Weak evidence
30

Create a Fair Infographic Plan

Plan a fair data infographic about screen time in Australia.

Where would you source your data? ___

What graph type would you use and why? ___

What information would you include to show it's trustworthy? ___

31

The SIFT Method

Use SIFT (Stop, Investigate, Find better sources, Trace claims) to evaluate a statistic.

Statistic you found: ___

Stop — what is your first reaction? ___

Investigate the source: ___

Find a second source that confirms or denies: ___

Trace the original claim back to: ___

32

Percentage Change

Circle the correct interpretation.

Sales went from 100 to 120. Change?

20% increase
120% increase
20 increase

Prices fell from 200 to 150. Change?

25% decrease
50% decrease
75% decrease

Enrolment doubled from 50 to 100. Change?

100% increase
50% increase
200% increase

From 80 to 60. Change?

25% decrease
20% decrease
33% decrease
33

Statistical Literacy Reflection

Reflect on what you have learned about reading statistics critically.

Three questions I now ask when I see a statistic: ___

A real example where statistics were used misleadingly: ___

Why is statistical literacy an important life skill? ___

34

Match the Logical Fallacy

Match each fallacy name to its description.

Cherry picking
False causation
Hasty generalisation
Anchoring
Small sample used for big claims
First number heard influences judgement
Ignoring data that doesn't fit
Two things happen together, so one causes the other
35

Percentage vs Absolute Numbers

Decide whether percentage or absolute number is more meaningful.

Disease doubled — from 2 cases to 4 cases. More alarming: 100% increase or 2 extra cases? ___

Tax cut: 1% — means $10 to a minimum wage worker, $10,000 to a millionaire. What does % hide? ___

Write your own example where absolute numbers matter more than percentages: ___

36

Write a Data Story

Use real or invented data to write a short, fair news article.

Topic: ___

Data source: ___

Key statistic: ___

Fair interpretation: ___

What other information would help readers understand? ___